基于互信息的多基地InISAR成像融合新方法  

A Novel Multiple InSAR Image Fusion Method Based on Mutual Information

作　　者：王秀友[1] 范建中[1] 徐冬青[1] 刘华明[1] 

机构地区：[1]阜阳师范学院计算机与信息工程学院,安徽阜阳236037

出　　处：《控制工程》2017年第11期2249-2255,共7页Control Engineering of China

基　　金：安徽省高校优秀青年骨干人才项目(gxfx2017072);安徽省自然科学基金(1708085MF155);阜阳师范学院青年人才基金重点项目(rcxm201706)

摘　　要：针对当前多基地雷达系统中传统算法无法利用各自孤立的信息进行目标有效重建的现状,提出了一种利用互信息(Mutual Information)进行目标干涉配准及融合的研究方法。首先,对干涉In ISAR回波模型及特点进行了分析。其次,提出了利用三维目标的二维投影进行仿射变换寻求最优互信息匹配参数,当两幅图像配准时,互信息达到最大。最后,将各个投影平面内融合好的二维图像重构目标的三维空间分布。仿真结果明确表明,融合后的三维分布比单独In ISAR系统可以更好的展现飞机原始模型的分布,目标更多的特征得以表现。与传统方法进行对比,其成像图像质量有了明显提升,验证了算法的有效性。In this paper we aim at solving the object reconstruction problem in the multi-base radar system, where the traditional algorithm cannot use isolated information to reconstruct the object. We propose a novel algorithm to use mutual information to register the object and fuse images. First, we analyze the interference InISAR echo model. Second, we use the 2D projection of 3D object to perform affine transformation to search for optimal parameters. The mutual information reaches its maximum when two images are registered. Third, the object is reconstructed in 3D space using the 2D image fused in the projection space. Simulation results show that the 3D airplane model distribution after fusion has outperformed the single InISAR system and more object features are demonstrated. Compared with the traditional method, the imaging quality has significant advantages, and the proposed method is proved to be effective.

关 键 词：图像融合 干涉ISAR图像 互信息 配准 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]